(IJACSA) International Journal of Advanced Computer Science and Applications,

Vol. 13, No. 3, 2022

Design and Implementation of a Low-cost CO2

Monitoring and Control System Prototype to
Optimize Ventilation Levels in Closed Spaces

Ramcés Cavallini-Rodrı́guez Jesús Espinoza-Valera Carlos Sotomayor-Beltran

Facultad de Ingenierı́a Facultad de Ingenierı́a Facultad de Ingenierı́a
Universidad Tecnológica del Perú Universidad Tecnológica del Perú Universidad Tecnológica del Perú
Lima, Peru Lima, Peru Lima, Peru

Abstract—High concentrations of CO2 levels are significantly
present in closed environments that do not have proper venti-
lation. Such high concentrations generate negative health con-
sequences such as dizziness, headaches and various respiratory
problems. For this reason, the design and implementation of
a low-cost CO2 monitoring and control prototype is proposed
to optimize ventilation levels in closed spaces. The parameters
that the proposed device measures are concentration of carbon
dioxide, humidity and temperature. A digital PID controller was
implemented, with the use of C++ programming language and
an exhaust fan to stabilize carbon dioxide levels within a closed
space. The aforementioned parameters can be viewed in two
ways: The first way is locally through a LCD screen and LED
indicators, and the second one, remotely using the free Arduino
IoT Cloud platform. The closed environment was emulated using
a cardboard box and in the tests it was obtained that the
prototype manages to keep the CO2 concentration levels below the
established limit. However, this can be further improved by using
more precise sensors for more accurate results. It is expected
that this model can be successfully scaled to closed spaces such
as classrooms and offices.
Keywords—CO2 Monitoring; IoT; Low-cost Indoor Ventilation
System; NodeMCU; Open Source Software
I. I NTRODUCTION
through an LCD screen and warns us with an audible alarm
when the maximum levels established are exceeded [4]. The
aforementioned studies [2], [3], [4] defend the importance of
monitoring the levels of polluting gases, such as CO2 , existing
inside a closed space. This, with the aim of analyzing indica-
tors that allow reducing the spread of diseases in the short
and long term. The tendency of the authors in this regard is to
develop portable prototypes that determine the concentration of
harmful agents to health within the environment using low-cost
sensors that provide information in real time, which seeks to
maintain adequate ventilation levels below the recommended
limit.
CO2 is a gas that is released by human respiration and
this is adequate to measure ventilation levels [5]. Because
people exhale predictable levels of CO2 , a direct connection
to air pollution can be made [3]. This parameter is extremely
important, since high concentrations of CO2 have very harmful
effects on health [6]. On the other hand, the measurement
of other parameters of air quality, such as particulate matter,
relative humidity and temperature, are very important as well.
Environmental parameters, temperature and humidity, also
influence people’s well-being [7], [8] and these are necessary
to have a correct indoor air quality index [2].

Currently, outdoor air pollution has increased in all coun-
tries due to the growth of automotive fleets and industrial
companies. This affects indoor environments with a significant
increase in CO2 levels in a proportion between 2 to 5 times
more than outdoors [1]. An example of this is reflected inside
different environments such as classrooms, offices, bedrooms,
etc; where the CO2 concentration has values that exceed
the maximum levels recommended by international standards.
These places do not have an adequate mechanical ventilation
system to dissipate the polluting gases that are generated
inside, and even in many cases the doors and windows are kept
closed. Likewise, despite the fact that they have been designed
to accommodate a limited number of people, the maximum
capacity allowed is exceeded on a daily basis.
Different CO2 monitoring systems have been developed
for closed spaces that do not directly control this parameter.
A CO2 prediction model has been developed based on the
history of measurements [2]. A user alert system has been also
proposed by means of text messages to the cell phone and an
email notification [3]. The information obtained is even shown
www.ijacsa.thesai.org
Some devices that has been previously designed to measure
different parameters from air quality have shown data in real
time through the uses of displays. For example, an open
source web platform was developed to display the information
provided by the sensors using a MySQL server for data storage
[7]. ThingSpeak was also used, which is a free web platform
for the collection and representation of the measured variables
[3]. On the other hand, a mobile phone application called
Dataplicity was used to remotely access the pollutant indicators
[9], whereas in another study only a 16x2 LCD screen was
used to fulfill the same purpose [10].
When considering the methodology used by some studies
to develop devices that measure air quality, diverse ones were
followed. For instance, a quantitative experimental investiga-
tion was applied in 10 schools and in 2 training centers for 8
days and between 5 to 10 hours a day taking measurements
from different air quality sensors [11]. A similar methodolog-
ical approach is applied using pollutant weighted calculation
tools and the test is performed in 2 environments. The first is
a laboratory at the University of Warwick (England) during 1
529 | P a g e
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 13, No. 3, 2022

hour of classes with a number of 100 students. The second

environment is a kitchen of the same school, with a single
person preparing the food for 1 hour [1]. On the other hand,
in one study [8] a qualitative methodology based on surveys
was used as an instrument to understand the experience of
discomfort and evaluate the environment from the participants
own point of view. Then, a statistical analysis was performed
based on census data for 8 hours a day in 4 scenarios. Similarly,
in another study [12] a systemic perspective of the model Fig. 1. CCS 811 Sensor.
designed with a smoothing algorithm to reduce sensor errors
caused by power failures and a data aggregation algorithm that
minimizes network traffic and saves time was developed. In a
research carried out in India [10], empirical methods were used
to know the level of gases harmful to health using the designed
prototype, but they did not specify the place or conditions
where it was used.
A research carried out in Spain [11], presents as a result
a very precise device whose data had a value R2 = 0.902 in
a simple linear regression test, and S = 143 and p = 0.1348
in the Moses test comparing it with the commercial Perfect
Prime CO2000 device. Similarly, researchers in Portugal [13] Fig. 2. DHT11 Sensor.
presented a very reliable proposal, showing sensors with R2
values between 0.73 and 0.87, obtained from linear regression
tests. Another work in India [3] explained a system with good
sensitivity, which although it was not compared with another 1) Acquisition Stage: Two types of sensors were imple-
device, had a sensor which was experimentally calibrated for mented in this stage.
24 hours.
• CCS 811 Sensor (Fig. 1): It is an air quality sensor
Some studies have concluded that portable devices are that allows us to obtain the equivalent concentration of
reliable to measure the quality of the air inside closed en- CO2 from the measurement of Total Volatile Organic
vironments and also confirm that the use of IoT sensors is Compounds (VOC) in a very precise way. This sensor
favorable for these equipments. For example, the use of IoT was chosen as a main parameter for CO2 measurement
technology simplifies access to the information provided by due to its low economic value compared to other types
the different sensors installed in the environment in order to of air quality sensor on the market that cost up to 6
have a better quality of air [14]. With the information collected times more.
by sensors interconnected by means of IoT technology, it
was possible to develop an anticipated model based on CO2 • DHT11 sensor (Fig. 2): It is a sensor that measures
concentrations to improve ventilation at any time [2]. On the temperature and humidity parameters. It is a digital
other hand, Interaction with a wireless sensor network that type that can be powered with 3.5 to 5V. Its tem-
provides real-time pollutant gas indices is also affordable for perature measurement range is 0 to 50◦ C, with an
everyone. Mounting a network of wireless sensors is easier and accuracy of 2◦ C. It allows to measure the relative
more efficient and naturally, it reduces the amount of wiring humidity of 20-80% with a maximum sampling period
used from the end devices to the central monitoring station of 1 second.
[15]. Finally, the Raspberry Pi development board facilitates
the connection with other IoT sensors and not only with analog 2) Control Stage: It is made up solely of the NodeMCU
or digital sensors to collect information on the contamination module.
rates in a closed area [12].
• NodeMCU v3 module (Fig. 3): This module has an
In view of this global problem, which is detrimental to ESP8266 processor at a speed of 80 to 160 MHz. It
people’s health, of not having an adequate CO2 monitoring has 1 analog port, 17 general purpose ports, 4 PWM
and control system, it is necessary to design and implement outputs and allows I2C, UART, SPI type communica-
one that optimizes the level of ventilation in closed spaces [7]. tion and Wifi. This device was chosen because it is a
very popular and cheap module to carry out electronic
projects that are synchronized with a free service in
II. M ETHODOLOGY the cloud every second, due to its low consumption
and high performance.
A. Electronic Components
3) Communication Stage: It will be done wirelessly
The prototype is divided into five operating stages and through the Wifi technology of the NodeMCU module. This
uses the following electronic components, with the aim of communication was chosen because this technology is incor-
achieving a low-cost system that can be implemented in closed porated into the NodeMCU development board and that it has
environments. a free server called Arduino IoT Cloud to be able to view the
www.ijacsa.thesai.org 530 | P a g e
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Fig. 3. NodeMCU v3 Module.
Fig. 4. 16X2 LCD + I2C Module.
KPI control indicators remotely, both on the web and on the
cell phone.
4) Display Stage: It is made up of a 16X2 LCD screen
with I2C module and 3 LEDs.
• 16X2 LCD Screen + I2C Module (Fig. 4): This
component is a small screen that allows information to
be displayed. It has 2 rows and in each row you have
the possibility of observing up to 16 alphanumeric
characters.
• It consists of 3 light emitting diodes of green, yellow
and red respectively. They will allow to show visually
based on a color of the led a range of the CO2 sensor
to indicate if the ppm values are low, medium or high,
respectively.
5) Actuator Stage: It is made up of 1 pre-actuator and 1
actuator.
• AC 220V / 2A digital dimmer (Fig. 5): Provides alter-
nating current voltage regulation in order to achieve
different levels of voltages that are located within the
voltage range, while varying the power delivered to
the load. This is controlled by a PWM signal. : This
device was chosen as a pre-actuator to gradually vary
the speed of the extractor motor, due to the effect of
the PID controller established to maintain the optimal
CO2 concentration at an approximate value of 1000
ppm.
• 220V Extracting fan (Fig. 6): It is that device that ex-
tracts the air through its rotating blades in a direction
from inside to outside of an enclosure. It works with
voltages of 220-240 VAC and at frequencies of 50 and
60 Hz. In addition, they have a protection impedance.
This device was chosen to extract the concentrated
gas into the environment when CO2 levels exceed a
dangerous value for human health and thus dissipate
it.
A diagram of the five stages working together in the design
of the prototype is show in Fig. 7.
www.ijacsa.thesai.org
Vol. 13, No. 3, 2022
Fig. 5. AC Digital Dimmer.
Fig. 6. 220V Air Extractor.
B. Mechanical Structure
Fig. 8 shows the design of the structure through the CAD
Solidworks 2020 program, with general dimensions of 102 x
50 x 110 mm and 2 mm thick, which will contain the system
components (LCD screen, sensors and plate development) so
that they fit perfectly. After that, a 3D printing can be made
using Acrylonitrile Butadiene Styrene (ABS) as a material, to
protect the elements of the system. ABS material was chosen
because it has sufficient protection against dust and splashes
of water.
The front, side, rear and bottom views are seen in Fig. 9,
Fig. 10, Fig. 11, and Fig. 12, respectively.
C. Software
The programming of the NodeMCU v3 module was done
through the Arduino IDE platform, since the module has
compatibility with that development environment. The goal
is to keep the CO2 concentration at 1000 ppm through the
use of a digital PID controller. Likewise, the temperature and
humidity parameters were considered in order to compensate
for the variations in CO2 .
First, the readings of the temperature and humidity sensor
were taken to be able to introduce them to the CO2 sensor
and improve the variation of the reading of its concentration.
The PID controller is activated when the carbon dioxide
concentration exceeds the value of 1500 ppm, because from
this concentration it begins to be very harmful to health. With
the establishment of the desired level of CO2 and the correct
tuning of the controller, the output value of the control signal is
calculated. The power that will be delivered to the air extractor
is the output of the PID controller, therefore a restriction of
its maximum value was established at 100 and minimum at 0.
This rectified output signal will be sent to the dimmer,
which will regulate the power that will be sent to the extractor.
Fina1glly, when the carbon dioxide concentration is less than
900 ppm, the PID controller will stop working, since the
531 | P a g e
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 13, No. 3, 2022

Fig. 7. Interconnection of the 5 Stages of the System.

Fig. 8. Design View of the Exterior Structure.

Fig. 9. Front View of the Exterior Structure.

D. Data Transmission to the Cloud

ventilation level of the place will be safe. The values of The measurements from the sensors are sent to the Arduino
temperature, humidity and CO2 concentration will be sent to IoT Cloud platform. This platform is free, easy to use and all
the cloud, displayed on the LCD screen and there will be 3 standards are open source. In addition to being compatible with
LED indicators: green when the CO2 concentration is less than all Arduino boards, it also has compatibility with the ESP8266
1000 ppm, yellow when it is between the range 1000 - 1500 microcontroller, which is used by the NodeMCU module. This
ppm and red when higher than 1500 ppm. Fig. 13 shows the platform uses the MQTT protocol for data transmission and
logic used for programming the entire system. allows other interaction methods such as command line tools,
www.ijacsa.thesai.org 532 | P a g e
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Fig. 10. Side View of the Exterior Structure.
Fig. 11. Rear View of the Exterior Structure.
Fig. 12. Bottom View of the Exterior Structure.
JavaScript, and WebSockets. It even has a data storage of up
to 2 days.
In this platform, 3 variables corresponding to each of the
parameters that are obtained from the sensors were created
and it was established that they are updated when their value
changes. An interface called “CO2 Monitoring and Control
System” was also designed and added a percentage progress
www.ijacsa.thesai.org
Vol. 13, No. 3, 2022
Fig. 13. Flow Diagram of System Operation.
indicator for the humidity value, a nominal progress indicator
for the temperature value and a trend graph for the value of
CO2 concentration in ppm, as shown in Fig. 14.
III. R ESULTS AND D ISCUSSION
For the tests of the designed prototype, in Fig. 15 we
have emulated a closed environment based on a cardboard
box measuring 41x55x38 cm, in which 27 small holes were
made, with an approximate diameter of 2 cm, on both sides
and in the back of the box so that the air can circulate when
the extractor is turned on, since otherwise there would not be
an exchange of the flow of clean and polluted air inside the
box. The cardboard box was chosen because of its low price
and that it allowed to accumulate carbon dioxide in such a
way as it would be done in a real closed space. However,
the use of other materials such as resin or glass, as in the
case of the research carried out in Algeria [16], would have
allowed allow to observe the total operation of the system and
to obtain a more natural behavior of the CO2 due to a better
permeability. Also, a plastic box with IP56 protection could
be used to protect the system components and be the external
structure of the prototype, as was done in England [17] and
Indonesia [7].
As can be seen in Fig. 16, a hole with a diameter of 11.5
cm was made in the front of the box so that the air extractor
can be installed outside of it. Likewise, 4 small holes were
533 | P a g e
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Fig. 14. KPI Indicators of the Cloud Interface.
Fig. 15. View of the Cardboard Box where the Prototype was Tested.
made in the corners to hold the extractor with screws and
washers. This extractor will circulate the air from the inside to
the outside of the box when the CO2 levels are not within the
allowed ranges. One type of exhaust fan that should be used in
a real room is the AC180 model from the Broan brand, which
works at 200 VAC 50Hz. In addition, it consumes 18W of
power with an extraction capacity of 163 m3 / h and a noise
level of 45 Db, being indicated for residential environments
such as rooms, offices and classrooms. In a study conducted
in a city in India (Durgapur), 6 conventional ”cooler” type
air extractors were used to ventilate the CO2 produced in the
small environment [8]. Also, in an office space in the same
country, generic mechanical ventilation was used, producing a
lot of noise [18], as well as a conventional extractor that was
activated when the carbon dioxide limits were exceeded [19].
It can be seen in Fig. 17 that inside the box the breadboard
was placed with the different components that intervene in the
system, such as the NodeMCU development board, the CO2
sensor, the temperature and humidity sensor, the LEDs and the
screen. LCD coupled to its I2C module. The LCD screen and
www.ijacsa.thesai.org
Vol. 13, No. 3, 2022
Fig. 16. Hole for Extractor Installation.
Fig. 17. Distribution of the Components within the Cardboard Box.
its module are powered with 5V DC through an adapted cell
phone charger, the Node MCU, through a microUSB cable
connected to the laptop and the other components use 3V
DC provided by the development board. In the middle part
of the box it can be seen the dimmer that will function as a
pre-actuator, regulating the power delivered and enabling the
connection between the control stage led by the NodeMCU
and the power stage where the air extractor is located.
Fig. 18 shows the local display stage made up of the
16x2 LCD screen, which provides us with information about
the values obtained from the CO2 concentration through the
CSS811 sensor in ppm measurements, and the temperature and
relative humidity through the DHT11 sensor in ◦ C and %
measurements, respectively.
In Fig. 19, we roughly mimic the CO2 exhaled by people
through smoke generated by burning paper in order to perform
the tests. A quarter of rolled paper is set on fire and immedi-
ately placed inside the box on a metal container, to finally close
it. The exhaust fan would be located in a real environment at
least 2.3 meters from the ground, embedded in the wall. In the
same way, the prototype that contains the CO2 sensor must be
mounted on the wall with the correct orientation, because CO2
534 | P a g e
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Fig. 18. Local Display Stage.
Fig. 19. CO2 Generation within the Cardboard Box.
is denser than air, it tends to fall to the ground and with this
it would absorb the CCS 811 sensor located at the interior
of the structure, which has ventilation grills that will allow
the access of this gas. A research work has used the MQ135
sensor to measure the CO2 concentration and have located this
at ground level [12]. Also, in a study in India, the MQ2 gas
sensor is used interconnected with Arduino and installed on
top of a desk [20]. Similarly in another work, an all-in-one
prototype is designed, which is installed on the wall [17].
Fig. 20 shows the result of test 1 represented with a trend
graph of CO2 concentration (ppm) vs. time (seconds), sampled
in 5 second intervals. At the beginning, an establishment is
observed at approximately 8000 ppm, since it was the initial
moment where the smoke was generated and accumulated.
Also, it should be noted that an important factor is the size of
www.ijacsa.thesai.org
Vol. 13, No. 3, 2022
Fig. 20. Test Result 1.
Fig. 21. Test Result 2.
the environment since, being a small box, the smoke generates
a CO2 concentration inside the box much higher than in a
natural and conventional scenario. After some time, a decrease
in the CO2 concentration occurs as a result of the action of
the air extractor, thus achieving an approximate establishment
of 1000 ppm after 800 seconds.
In Fig. 21 the result of test 2 is shown, in addition to
the initial establishment period at 8000 ppm, some downward
peaks are observed since being a non-natural environment and
only simulated, the generated CO2 conditions are not constant,
as if they would be in a traditional environment. Despite this,
in general a similar behavior is observed with respect to the
results of test 1, stabilizing at approximately 1000 ppm after
800 seconds, thus generating a clean and safe environment. In
both tests carried out, the correct extraction of carbon dioxide
inside the box can be observed, through the exponential decay
of its concentration seen in both graphs. This behavior is
repeated in some studies carried out in Italy [21] and the
United States [22] inside classrooms, varying the number of
people and ventilation elements. This was also observed in an
investigation conducted in two laboratories in Indonesia [23].
In Fig. 22 the remote viewing stage is shown through
the free Arduino IoT Cloud platform that can be monitored
through the web or its free software for Android and iOS.
535 | P a g e
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 13, No. 3, 2022

Fig. 22. Remote Viewing Stage.

In either of these two media, KPI indicators for humidity,
temperature and CO2 concentration are displayed in real time
with a history of up to 2 days.
IV. C ONCLUSION
In the present research work, a monitoring and control sys-
tem of CO2 concentration in closed rooms has been developed
using low-cost components and a free software application
to view the data remotely through smartphones, laptops and
desktop computers. The results demonstrate its effectiveness
compared to other studies, since after approximately 800 sec-
onds of having detected the high levels of CO2 concentration, it
reaches adequate levels, less than 1000 ppm, thus producing a
healthy area for the exchange of air of the people who inhabit
it. The finished prototype would be located on the side of
the room for the correct detection of the parameters described
above.
As a future work, it is suggested that the development of
the prototype be located within a structure with a degree of
protection IP (Ingress Protection), with the aim of resisting
dust and humidity found in the environment. Likewise, it is
recommended to use CO2 sensors of the NDIR (Non Dis-
persive Infrared Detector) type, since they are more accurate
for gas detection, easier to calibrate and have minimal error
percentages. Take into account the number of occupants and
the volume of the room, since the precision of the sensors to be
used and the number of air extractors to install depends on this.
Finally, to obtain accurate measurements it is recommended to
install the structure approximately 1 meter from the ground.
R EFERENCES
[1] S. Esfahani, P. Rollins, J. P. Specht, M. Cole y J. W. Gardner,“Smart City
Battery Operated IoT Based Indoor Air Quality Monitoring System”, en
2020 IEEE SENSORS, Rotterdam, 25–28 october 2020. 2020.
[2] R. Xiahou, J. Yi, L. He, W. He y T. Huang, “Indoor air monitoring system
based on internet of things and its prediction mode’l’, Proceedings of

www.ijacsa.thesai.org 536 | P a g e
 (IJACSA) International Journal of Advanced Computer Science and Applications,
the International Conference on Industrial Control Network and System
Engineering Research (ICNSER), vol. 2019, pp. 58–63, 2019
[3] R. K. Kodali, S. Pathuri y S. C. Rajnarayanan, “Smart Indoor Air
Pollution Monitoring Station”, en 2020 International Conference on
Computer Communication and Informatics (ICCCI), Coimbatore, India,
22–24 january 2020. IEEE, 2020.
[4] C. Calvin Adrianto, Y. Calvinus y P. B. Mardjoko, “Design Of CO, CO2,
Temperature, Humidity, And Weather Monitoring System Based On
Internet of Things (IoT) & Android”, IOP Conference Series: Materials
Science and Engineering, vol. 1007, p. 012175, december 2020
[5] D. Thomas, B. Mistry, S. Snow y M. C. Schraefel, “Indoor Air Quality
Monitoring (IAQ): A Low-Cost Alternative to CO2 Monitoring in
Comparison to an Industry Standard Device”, Advances in Intelligent
Systems and Computing, vol. 858, n. 1, pp. 1010–1027, 2018
[6] G. Marques y R. Pitarma, “IAQ Evaluation Using an IoT CO2 Monitor-
ing System for Enhanced Living Environments”, en Advances in Intelli-
gent Systems and Computing, Cham: Springer International Publishing,
2018, pp. 1169–1177
[7] A. A. Hapsari, A. I. Hajamydeen, D. J. Vresdian, M. Manfaluthy, L.
Prameswono y E. Yusuf, “Real Time Indoor Air Quality Monitoring
System Based on IoT using MQTT and Wireless Sensor Network”, en
2019 IEEE 6th International Conference on Engineering Technologies
and Applied Sciences (ICETAS), Kuala Lumpur, Malaysia, 20–21 de-
cember 2019. IEEE, 2019.
[8] P. K. Sharma, B. Poddar, S. Dey, S. Nandi, T. De, M. Saha, S. Mondal
y S. Saha, “On detecting acceptable air contamination in classrooms
using low cost sensors”, 9th International Conference on Communication
Systems and Networks (COMSNETS), vol. 2017, pp. 484–487, 2017
[9] D. Suriano, G. Cassano y M. Penza, “Design and Development of a
Flexible, Plug-and-Play, Cost-Effective Tool for on-Field Evaluation of
Gas Sensors”, Journal of Sensors, vol. 2020, pp. 1–20, august 2020.
[10] P. Satyanarayana y R. Narmadha, “Implementation of Wireless Sensor
Network Based Indoor Air Quality Monitoring System using GSM”,
International Journal of Engineering and Advanced Technology, vol. 8,
n. 6, pp. 4885–4889, august 2019.
[11] M. D. Fernández-Ramos, F. Moreno-Puche, P. Escobedo, P. A. Garcı́a-
López, L. F. Capitán-Vallvey y A. Martı́nez-Olmos, “Optical portable
instrument for the determination of CO2 in indoor environments”,
Talanta, vol. 208, p. 120387, february 2020.
[12] J.-Y. Kim, C.-H. Chu y S.-M. Shin, “ISSAQ: An Integrated Sensing
Systems for Real-Time Indoor Air Quality Monitoring”, IEEE Sensors
Journal, vol. 14, n. 12, pp. 4230–4244, december 2014.
[13] M. Lopes, J. Reis, A. P. Fernandes, D. Lopes, R. Luorenco, T. Nunes,
www.ijacsa.thesai.org
Vol. 13, No. 3, 2022
C. H. G. Faria, C. Borrego y A. I. Miranda, “Indoor air quality study
using low-cost sensors”, en AIR POLLUTION 2020, 8–10 june 2020.
Southampton UK: WIT Press, 2020.
[14] F. Tahsiin, L. Anggraeni, I. Chandra, R. A. Salam y H. Bethaningtyas,
“Analysis of Indoor Air QualityBased on Low-Cost Sensors”, Inter-
national Journal on Advanced Science, Engineering and Information
Technology, vol. 10, n. 6, pp. 2627–2633, december 2020.
[15] A. K. Kanal y T. Kovacshazy, “IoT Solution for Assessing the In-
door Air Quality of Educational Facilities”, en 2019 20th International
Carpathian Control Conference (ICCC), Krakow-Wieliczka, Poland,
26–29 may 2019. IEEE, 2019.
[16] A. Mellit, M. Benghanem, O. Herrak y A. Messalaoui, “Design of
a Novel Remote Monitoring System for Smart Greenhouses Using the
Internet of Things and Deep Convolutional Neural Networks”, Energies,
vol. 14, n. 16, p. 15, august 2021.
[17] T. Pietrucha,“Ability to Determine the Quality of Indoor Air in Class-
rooms without Sensors”, E3S Web of Conferences, vol. 17, p. 00073,
2017.
[18] S. N. Mabdeh, A. Al-Zghoul, T. Alradaideh, A. Bataineh y S. Ahmad,
“Simulation study for natural ventilation retrofitting techniques in ed-
ucational classrooms – A case study”, Heliyon, vol. 6, n. 10, october
2020
[19] O. A. Postolache, J. M. D. Pereira y P. M. B. S. Girao, “Smart Sensors
Network for Air Quality Monitoring Applications”, IEEE Transactions
on Instrumentation and Measurement, vol. 58, n.º 9, pp. 3253–3262,
september 2009
[20] K. Shahzad y B. Oelmann, “A comparative study of in-sensor processing
vs. raw data transmission using ZigBee, BLE and Wi-Fi for data intensive
monitoring applications”, en 2014 11th International Symposium on
Wireless Communications Systems (ISWCS), Barcelona, Spain, 26–29
august 2014.
[21] A. Zivelonghi y M. Lai, “Mitigating aerosol infection risk in school
buildings: the role of natural ventilation, volume, occupancy and CO2
monitoring”, Building and Environment, vol. 204, p. 13, October 2021.
[22] J. Allen, J. Spengler, E. Jones, J. Cedeno-Laurent, “5-
step guide to checking ventilation rates in classrooms”.
Reporte tecnico, Harvard. https://schools.forh.ealth.org/wp-
content/uploads/sites/19/2021/01/Harvard-Healthy-Buildings-program-
How-to-assess-classroom-ventilation-10-30-2020-EN R1.8.pdf, 2020.
[23] L. V. de Abreu-Harbich, V. L. A. Chaves y M. C. G. O. Brandstetter,
“Evaluation of strategies that improve the thermal comfort and energy
saving of a classroom of an institutional building in a tropical climate”,
Building and Environment, vol. 135, pp. 257–268, may 2018
537 | P a g e